Skip to contentSkip to conversation

mm: Properly document tail pages for compound pages

RFC PATCH Superseded Peter Xu 3 participants 5 messages Aug 10, 2023 – Aug 14, 2023

Subsystems (2)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Peter Xu

author
#0 lore raw

RFC PATCH 1/1 mm: Properly document tail pages for compound pages

Tail page struct reuse is over-comlicated. Not only because we have implicit uses of tail page fields (mapcounts, or private for thp swap support, etc., that we _may_ still use in the page structs, but not obvious the relationship between that and the folio definitions), but also because we have 32/64 bits layouts for struct page so it's unclear what we can use and what we cannot when trying to find a new spot in folio struct.

We also have tricks like page->mapping, where we can reuse only the tail page 1/2 but nothing more than tail page 2. It is all mostly hidden, until someone starts to read into a VM_BUG_ON_PAGE() of __split_huge_page_tail().

Let's document it clearly on what we can use and what we can't, with 100% explanations on each of them. Hopefully this will make:

(1) Any reader to know exactly what field is where and for what, the relationships between folio tail pages and struct page definitions,

(2) Any potential new fields to be added to a large folio, so we're clear which field one can still reuse (look for _reserved* ones).

This is assuming WORD is defined as sizeof(void *) on any archs, just like the other comment in struct page we already have.

One pitfall is I'll need to split part of the tail page 1 definition into 32/64 bits differently, that introduced some duplications on the fields. But hopefully that's worthwhile as it makes everything crystal clear. Not to mention that "pitfall" also brings a benefit that we can actually define fields in different order for 32/64 bits when we want.

---
 include/linux/mm_types.h | 76 +++++++++++++++++++++++++++++++++++-----
 1 file changed, 67 insertions(+), 9 deletions(-)
include/linux/mm_types.h +67 -9
index 291c05cacd48..3e40e1b9fec3 100644
@@ -313,41 +313,99 @@ struct folio {
 		};
 		struct page page;
 	};
+	/*
+	 * Some of the tail page fields (out of 8 WORDs for either 32/64
+	 * bits archs) may not be reused by the folio object because
+	 * they're already been used by the page struct:
+	 *
+	 * |-------+---------------|
+	 * | Index | Field         |
+	 * |-------+---------------|
+	 * |     0 | flag          |
+	 * |     1 | compound_head |
+	 * |     2 | N/A [0]       |
+	 * |     3 | mapping [1]   |
+	 * |     4 | N/A [0]       |
+	 * |     5 | private [2]   |
+	 * |     6 | mapcount      |
+	 * |     7 | N/A [0]       |
+	 * |-------+---------------|
+	 *
+	 * [0] "N/A" marks fields that are available to leverage for the
+	 *     large folio.
+	 *
+	 * [1] "mapping" field is only used for sanity check, see
+	 *     TAIL_MAPPING.  Still valid to use for tail pages 1/2.
+	 *     (for that, see __split_huge_page_tail()).
+	 *
+	 * [2] "private" field is used when THP_SWAP is on (disabled on 32
+	 *     bits, or on hugetlb folios) .
+	 */
 	union {
 		struct {
+	/* WORD 0-1: not valid to reuse */
 			unsigned long _flags_1;
 			unsigned long _head_1;
-	/* public: */
+	/* WORD 2 */
 			unsigned char _folio_dtor;
 			unsigned char _folio_order;
+			unsigned char _holes_1[2];
+#ifdef CONFIG_64BIT
 			atomic_t _entire_mapcount;
+	/* WORD 3 */
 			atomic_t _nr_pages_mapped;
 			atomic_t _pincount;
-#ifdef CONFIG_64BIT
+	/* WORD 4 */
 			unsigned int _folio_nr_pages;
+			unsigned int _reserved_1_1;
+	/* WORD 5-6: not valid to reuse */
+			unsigned long _used_1_2[2];
+	/* WORD 7 */
+			unsigned long _reserved_1_2;
+#else
+	/* WORD 3 */
+			atomic_t _entire_mapcount;
+	/* WORD 4 */
+			atomic_t _nr_pages_mapped;
+	/* WORD 5: only valid for 32bits */
+			atomic_t _pincount;
+	/* WORD 6: not valid to reuse */
+			unsigned long _used_1_2;
+	/* WORD 7 */
+			unsigned long _reserved_1;
 #endif
-	/* private: the union with struct page is transitional */
 		};
+	/* private: the union with struct page is transitional */
 		struct page __page_1;
 	};
 	union {
 		struct {
+	/* WORD 0-1: not valid to reuse */
 			unsigned long _flags_2;
 			unsigned long _head_2;
-	/* public: */
+	/* WORD 2-5 */
 			void *_hugetlb_subpool;
 			void *_hugetlb_cgroup;
 			void *_hugetlb_cgroup_rsvd;
 			void *_hugetlb_hwpoison;
-	/* private: the union with struct page is transitional */
+	/* WORD 6: not valid to reuse */
+			unsigned long _used_2_2;
+	/* WORD 7: */
+			unsigned long _reserved_2_1;
 		};
 		struct {
-			unsigned long _flags_2a;
-			unsigned long _head_2a;
-	/* public: */
+	/* WORD 0-1: not valid to reuse */
+			unsigned long _used_2_3[2];
+	/* WORD 2-3: */
 			struct list_head _deferred_list;
-	/* private: the union with struct page is transitional */
+	/* WORD 4: */
+			unsigned long _reserved_2_2;
+	/* WORD 5-6: not valid to reuse */
+			unsigned long _used_2_4[2];
+	/* WORD 7: */
+			unsigned long _reserved_2_3;
 		};
+	/* private: the union with struct page is transitional */
 		struct page __page_2;
 	};
 };
Signature · 1 line
2.41.0
#1 lore raw
Quoted 6 lines · Peter Xu

Tail page struct reuse is over-comlicated. Not only because we have implicit uses of tail page fields (mapcounts, or private for thp swap support, etc., that we _may_ still use in the page structs, but not obvious the relationship between that and the folio definitions), but also because we have 32/64 bits layouts for struct page so it's unclear what we can use and what we cannot when trying to find a new spot in folio struct.

I do not like this patch.

Quoted 3 lines

We also have tricks like page->mapping, where we can reuse only the tail page 1/2 but nothing more than tail page 2. It is all mostly hidden, until someone starts to read into a VM_BUG_ON_PAGE() of __split_huge_page_tail().

We can change those BUG_ON if we want to reuse mapping in more tail pages. Ask!

Quoted 2 lines

Let's document it clearly on what we can use and what we can't, with 100% explanations on each of them. Hopefully this will make:

The explanations are still very page centric. I do not like the style of them, nor how you explain them.

Quoted 5 lines

One pitfall is I'll need to split part of the tail page 1 definition into 32/64 bits differently, that introduced some duplications on the fields. But hopefully that's worthwhile as it makes everything crystal clear. Not to mention that "pitfall" also brings a benefit that we can actually define fields in different order for 32/64 bits when we want.

No. This is going to ruin kernel-doc.

include/linux/mm_types.h struct folio { · line 317

Quoted 2 lines
include/linux/mm_types.h +2 -0
+	/*
+	 * Some of the tail page fields (out of 8 WORDs for either 32/64

There's your first mistake; struct page is not necessarily 8 WORDs. You've got 7 words for sure, then on 32-bit you have 8 because atomic_t is word-sized. memcg_data might be the 9th word, virtual could be the tenth, two awful kmsan intrusions could bring it to twelve, and last_cpupid could bring it to thirteen.

Sure, it's 8 words on x86-64 with CONFIG_MEMCG enabled. But that's just your system.

Quoted 4 lines
include/linux/mm_types.h +4 -0
+	 * bits archs) may not be reused by the folio object because
+	 * they're already been used by the page struct:
+	 *
+	 * |-------+---------------|

include/linux/mm_types.h struct folio { · line 331

Quoted 10 lines
include/linux/mm_types.h +10 -0
+	 * | Index | Field         |
+	 * |-------+---------------|
+	 * |     0 | flag          |
+	 * |     1 | compound_head |
+	 * |     2 | N/A [0]       |
+	 * |     3 | mapping [1]   |
+	 * |     4 | N/A [0]       |
+	 * |     5 | private [2]   |
+	 * |     6 | mapcount      |
+	 * |     7 | N/A [0]       |

This is wrong. You mustn't reuse refcount. refcount must remain 0 on all tail pages. And you can't use anything after refcount, because it's all optional on various configurations.

include/linux/mm_types.h struct folio { · line 335

Quoted 4 lines
include/linux/mm_types.h +4 -0
+	 * |-------+---------------|
+	 *
+	 * [0] "N/A" marks fields that are available to leverage for the
+	 *     large folio.

N/A is a bad way to say this. "Free" or "Available" would be better.

include/linux/mm_types.h struct folio { · line 339

Quoted 3 lines
include/linux/mm_types.h +3 -0
+	 * [1] "mapping" field is only used for sanity check, see
+	 *     TAIL_MAPPING.  Still valid to use for tail pages 1/2.
+	 *     (for that, see __split_huge_page_tail()).

No, definitely wrong to document this.

include/linux/mm_types.h struct folio { · line 342

Quoted 2 lines
include/linux/mm_types.h +2 -0
+	 * [2] "private" field is used when THP_SWAP is on (disabled on 32
+	 *     bits, or on hugetlb folios) .

Ugh, this needs to be fixed, not documented. If you really must document it, at least say that this needs to be fixed.

include/linux/mm_types.h struct folio { · line 346

Quoted 4 lines
include/linux/mm_types.h +2 -0
+	 */
 	union {
 		struct {
+	/* WORD 0-1: not valid to reuse */

... so now you're repeating all the information you provided above?

include/linux/mm_types.h struct folio { · line 320

Quoted 3 lines
include/linux/mm_types.h +0 -1
 			unsigned long _flags_1;
 			unsigned long _head_1;
-	/* public: */

... did you check kernel-doc after removing this?

include/linux/mm_types.h struct folio { · line 352

Quoted 4 lines
include/linux/mm_types.h +2 -0
+	/* WORD 2 */
 			unsigned char _folio_dtor;
 			unsigned char _folio_order;
+			unsigned char _holes_1[2];

No. If you need to search for holes, use pahole.

Quoted 18 lines
include/linux/mm_types.h +13 -1
+#ifdef CONFIG_64BIT
 			atomic_t _entire_mapcount;
+	/* WORD 3 */
 			atomic_t _nr_pages_mapped;
 			atomic_t _pincount;
-#ifdef CONFIG_64BIT
+	/* WORD 4 */
 			unsigned int _folio_nr_pages;
+			unsigned int _reserved_1_1;
+	/* WORD 5-6: not valid to reuse */
+			unsigned long _used_1_2[2];
+	/* WORD 7 */
+			unsigned long _reserved_1_2;
+#else
+	/* WORD 3 */
+			atomic_t _entire_mapcount;
+	/* WORD 4 */
+			atomic_t _nr_pages_mapped;

include/linux/mm_types.h struct folio { · line 378

Quoted 10 lines
include/linux/mm_types.h +7 -1
+	/* WORD 5: only valid for 32bits */
+			atomic_t _pincount;
+	/* WORD 6: not valid to reuse */
+			unsigned long _used_1_2;
+	/* WORD 7 */
+			unsigned long _reserved_1;
 #endif
-	/* private: the union with struct page is transitional */
 		};
+	/* private: the union with struct page is transitional */

You don't understand why I did it like this. Again, you have to build the kernel-doc and you'll see why the private: and public: markers are where they are.

There was even a thread on it, a year or two ago, where I outlined the various tradeoffs between complexity of the output and the complexity of the input.

author
#2 lore raw
Quoted 23 lines · Matthew Wilcox
Quoted 6 lines · Peter Xu

Tail page struct reuse is over-comlicated. Not only because we have implicit uses of tail page fields (mapcounts, or private for thp swap support, etc., that we _may_ still use in the page structs, but not obvious the relationship between that and the folio definitions), but also because we have 32/64 bits layouts for struct page so it's unclear what we can use and what we cannot when trying to find a new spot in folio struct.

I do not like this patch.

Quoted 3 lines

We also have tricks like page->mapping, where we can reuse only the tail page 1/2 but nothing more than tail page 2. It is all mostly hidden, until someone starts to read into a VM_BUG_ON_PAGE() of __split_huge_page_tail().

We can change those BUG_ON if we want to reuse mapping in more tail pages. Ask!

Quoted 2 lines

Let's document it clearly on what we can use and what we can't, with 100% explanations on each of them. Hopefully this will make:

The explanations are still very page centric. I do not like the style of them, nor how you explain them.

Quoted 7 lines
Quoted 5 lines

One pitfall is I'll need to split part of the tail page 1 definition into 32/64 bits differently, that introduced some duplications on the fields. But hopefully that's worthwhile as it makes everything crystal clear. Not to mention that "pitfall" also brings a benefit that we can actually define fields in different order for 32/64 bits when we want.

No. This is going to ruin kernel-doc.

My fault. I'll make sure kernel-doc will keep looking sane.

It'll mean that my below trick on separating for 32/64 bits with duplicated fields will stop working, but I found some other way to represent this.

Quoted 10 lines

include/linux/mm_types.h struct folio { · line 317

Quoted 3 lines
include/linux/mm_types.h +2 -0
+	/*
+	 * Some of the tail page fields (out of 8 WORDs for either 32/64

There's your first mistake; struct page is not necessarily 8 WORDs. You've got 7 words for sure, then on 32-bit you have 8 because atomic_t is word-sized. memcg_data might be the 9th word, virtual could be the tenth, two awful kmsan intrusions could bring it to twelve, and last_cpupid could bring it to thirteen.

Quoted 2 lines

Sure, it's 8 words on x86-64 with CONFIG_MEMCG enabled. But that's just your system.

True. I messed up the mapcount/refcount position on 64 bits, sorry.

I hope atomic_t can always be assumed as 4 bytes though, and should hardly change trivially, so I can still rely upon.

Quoted 11 lines
Quoted 11 lines
include/linux/mm_types.h +10 -0
+	 * bits archs) may not be reused by the folio object because
+	 * they're already been used by the page struct:
+	 *
+	 * |-------+---------------|
+	 * | Index | Field         |
+	 * |-------+---------------|
+	 * |     0 | flag          |
+	 * |     1 | compound_head |
+	 * |     2 | N/A [0]       |
+	 * |     3 | mapping [1]   |
Quoted 8 lines

include/linux/mm_types.h struct folio { · line 331

Quoted 4 lines
include/linux/mm_types.h +4 -0
+	 * |     4 | N/A [0]       |
+	 * |     5 | private [2]   |
+	 * |     6 | mapcount      |
+	 * |     7 | N/A [0]       |

This is wrong. You mustn't reuse refcount. refcount must remain 0 on all tail pages. And you can't use anything after refcount, because it's all optional on various configurations.

I got the answer in the other thread. I still don't know a full list of things that we may require that, the only two things I got now are: (1) fast-gup over any-sized folio on speculative boosts of refs, and (2) pfn walkers like has_unmovable_pages().

I'll make an example here describing this, but if anyone can have a full list of why it must keep zero here I think it'll be great to document all of them, and this seems to be one of the best places to do it.

Quoted 7 lines

include/linux/mm_types.h struct folio { · line 335

Quoted 5 lines
include/linux/mm_types.h +4 -0
+	 * |-------+---------------|
+	 *
+	 * [0] "N/A" marks fields that are available to leverage for the
+	 *     large folio.

N/A is a bad way to say this. "Free" or "Available" would be better.

Indeed. I'll choose FREE.

Quoted 6 lines

include/linux/mm_types.h struct folio { · line 339

Quoted 4 lines
include/linux/mm_types.h +3 -0
+	 * [1] "mapping" field is only used for sanity check, see
+	 *     TAIL_MAPPING.  Still valid to use for tail pages 1/2.
+	 *     (for that, see __split_huge_page_tail()).

No, definitely wrong to document this.

I can document it in a "less detailed" way. I tend to introduce one macro TAIL_MAPPING_REUSED_MAX=2 and use it in __split_huge_page_tail().

Quoted 6 lines

include/linux/mm_types.h struct folio { · line 342

Quoted 3 lines
include/linux/mm_types.h +2 -0
+	 * [2] "private" field is used when THP_SWAP is on (disabled on 32
+	 *     bits, or on hugetlb folios) .

Ugh, this needs to be fixed, not documented. If you really must document it, at least say that this needs to be fixed.

Sure. Any further description on the problem and "a proper fix"? I will document that as "may need fixing" for now, but I'd try to reference more on it if I can understand better on the problem.

Quoted 7 lines

include/linux/mm_types.h struct folio { · line 346

Quoted 5 lines
include/linux/mm_types.h +2 -0
+	 */
 	union {
 		struct {
+	/* WORD 0-1: not valid to reuse */

... so now you're repeating all the information you provided above?

I'll try to remove many of them, but keep some so we read the fields easier, knowing the offset of the WORD. I'm still open to remove all of them if you think that's preferred.

Quoted 7 lines

include/linux/mm_types.h struct folio { · line 320

Quoted 4 lines
include/linux/mm_types.h +0 -1
 			unsigned long _flags_1;
 			unsigned long _head_1;
-	/* public: */

... did you check kernel-doc after removing this?

Quoted 6 lines

include/linux/mm_types.h struct folio { · line 352

Quoted 4 lines
include/linux/mm_types.h +2 -0
+	/* WORD 2 */
 			unsigned char _folio_dtor;
 			unsigned char _folio_order;
+			unsigned char _holes_1[2];

No. If you need to search for holes, use pahole.

Actually using _holes_* is definitely not right, because holes are literally free spaces that can be reused. Here IMHO I should describe it as two free bytes just like below, and that's one major goal of such a patch, to be clear on what can be still reused in tail pages.

Quoted 12 lines

include/linux/mm_types.h struct folio { · line 362

Quoted 12 lines
include/linux/mm_types.h +6 -1
+#ifdef CONFIG_64BIT
 			atomic_t _entire_mapcount;
+	/* WORD 3 */
 			atomic_t _nr_pages_mapped;
 			atomic_t _pincount;
-#ifdef CONFIG_64BIT
+	/* WORD 4 */
 			unsigned int _folio_nr_pages;
+			unsigned int _reserved_1_1;
+	/* WORD 5-6: not valid to reuse */
+			unsigned long _used_1_2[2];

(These "_used*" things will go away; I took another trick to represent only
"free" slots not "used", side effect of making kernel-doc happy)

Quoted 22 lines
Quoted 7 lines
include/linux/mm_types.h +7 -0
+	/* WORD 7 */
+			unsigned long _reserved_1_2;
+#else
+	/* WORD 3 */
+			atomic_t _entire_mapcount;
+	/* WORD 4 */
+			atomic_t _nr_pages_mapped;

include/linux/mm_types.h struct folio { · line 378

Quoted 10 lines
include/linux/mm_types.h +7 -1
+	/* WORD 5: only valid for 32bits */
+			atomic_t _pincount;
+	/* WORD 6: not valid to reuse */
+			unsigned long _used_1_2;
+	/* WORD 7 */
+			unsigned long _reserved_1;
 #endif
-	/* private: the union with struct page is transitional */
 		};
+	/* private: the union with struct page is transitional */

You don't understand why I did it like this. Again, you have to build the kernel-doc and you'll see why the private: and public: markers are where they are.

Quoted 3 lines

There was even a thread on it, a year or two ago, where I outlined the various tradeoffs between complexity of the output and the complexity of the input.

Links would be appreciated, at the meantime I'll send a rfcv2. I hope that addresses most of the issues above while still provide most of what I wanted to provide; not sure about this last one though before I get more information. Let me know if the whole idea just won't work, then I can always stop.

Signature · 1 line
Peter Xu
#3 lore raw
In reply to Peter Xu
kernel test robot noticed the following build errors:

Hi Peter,

[This is a private test report for your RFC patch.] kernel test robot noticed the following build errors:

[auto build test ERROR on akpm-mm/mm-everything]

url:    https://github.com/intel-lab-lkp/linux/commits/Peter-Xu/mm-Properly-document-tail-pages-for-compound-pages/20230811-045015
base:   https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git mm-everything
patch link:    https://lore.kernel.org/r/20230810204944.53471-1-peterx%40redhat.com
patch subject: [PATCH RFC] mm: Properly document tail pages for compound pages
config: mips-randconfig-r035-20230811 (https://download.01.org/0day-ci/archive/20230811/[email protected]/config)
compiler: mips64el-linux-gcc (GCC) 12.3.0
reproduce: (https://download.01.org/0day-ci/archive/20230811/[email protected]/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/

All errors (new ones prefixed by >>):

In file included from include/linux/bits.h:21,

                    from arch/mips/include/asm/cpu.h:12,
                    from arch/mips/include/asm/timex.h:16,
                    from include/linux/timex.h:67,
                    from include/linux/time32.h:13,
                    from include/linux/time.h:60,
                    from include/linux/compat.h:10,
                    from arch/mips/kernel/asm-offsets.c:12:
>> include/linux/build_bug.h:78:41: error: static assertion failed: "offsetof(struct folio, _flags_2) == offsetof(struct page, flags) + 2 * sizeof(struct page)"
      78 | #define __static_assert(expr, msg, ...) _Static_assert(expr, msg)
         |                                         ^~~~~~~~~~~~~~
   include/linux/build_bug.h:77:34: note: in expansion of macro '__static_assert'
      77 | #define static_assert(expr, ...) __static_assert(expr, ##__VA_ARGS__, #expr)
         |                                  ^~~~~~~~~~~~~~~
   include/linux/mm_types.h:434:9: note: in expansion of macro 'static_assert'
     434 |         static_assert(offsetof(struct folio, fl) ==                     \
         |         ^~~~~~~~~~~~~
   include/linux/mm_types.h:436:1: note: in expansion of macro 'FOLIO_MATCH'
     436 | FOLIO_MATCH(flags, _flags_2);
         | ^~~~~~~~~~~
>> include/linux/build_bug.h:78:41: error: static assertion failed: "offsetof(struct folio, _head_2) == offsetof(struct page, compound_head) + 2 * sizeof(struct page)"
      78 | #define __static_assert(expr, msg, ...) _Static_assert(expr, msg)
         |                                         ^~~~~~~~~~~~~~
   include/linux/build_bug.h:77:34: note: in expansion of macro '__static_assert'
      77 | #define static_assert(expr, ...) __static_assert(expr, ##__VA_ARGS__, #expr)
         |                                  ^~~~~~~~~~~~~~~
   include/linux/mm_types.h:434:9: note: in expansion of macro 'static_assert'
     434 |         static_assert(offsetof(struct folio, fl) ==                     \
         |         ^~~~~~~~~~~~~
   include/linux/mm_types.h:437:1: note: in expansion of macro 'FOLIO_MATCH'
     437 | FOLIO_MATCH(compound_head, _head_2);
         | ^~~~~~~~~~~

make[3]: *** [scripts/Makefile.build:116: arch/mips/kernel/asm-offsets.s] Error 1 shuffle=3596142661 make[3]: Target 'missing-syscalls' not remade because of errors.
make[2]: *** [arch/mips/Makefile:434: archprepare] Error 2 shuffle=3596142661 make[2]: Target 'prepare' not remade because of errors.
make[1]: *** [Makefile:234: __sub-make] Error 2 shuffle=3596142661
make[1]: Target 'prepare' not remade because of errors.
make: *** [Makefile:234: __sub-make] Error 2 shuffle=3596142661
make: Target 'prepare' not remade because of errors.

vim +78 include/linux/build_bug.h

bc6245e5efd70c Ian Abbott       2017-07-10  60  
6bab69c65013be Rasmus Villemoes 2019-03-07  61  /**
6bab69c65013be Rasmus Villemoes 2019-03-07  62   * static_assert - check integer constant expression at build time
6bab69c65013be Rasmus Villemoes 2019-03-07  63   *
6bab69c65013be Rasmus Villemoes 2019-03-07  64   * static_assert() is a wrapper for the C11 _Static_assert, with a
6bab69c65013be Rasmus Villemoes 2019-03-07  65   * little macro magic to make the message optional (defaulting to the
6bab69c65013be Rasmus Villemoes 2019-03-07  66   * stringification of the tested expression).
6bab69c65013be Rasmus Villemoes 2019-03-07  67   *
6bab69c65013be Rasmus Villemoes 2019-03-07  68   * Contrary to BUILD_BUG_ON(), static_assert() can be used at global
6bab69c65013be Rasmus Villemoes 2019-03-07  69   * scope, but requires the expression to be an integer constant
6bab69c65013be Rasmus Villemoes 2019-03-07  70   * expression (i.e., it is not enough that __builtin_constant_p() is
6bab69c65013be Rasmus Villemoes 2019-03-07  71   * true for expr).
6bab69c65013be Rasmus Villemoes 2019-03-07  72   *
6bab69c65013be Rasmus Villemoes 2019-03-07  73   * Also note that BUILD_BUG_ON() fails the build if the condition is
6bab69c65013be Rasmus Villemoes 2019-03-07  74   * true, while static_assert() fails the build if the expression is
6bab69c65013be Rasmus Villemoes 2019-03-07  75   * false.
6bab69c65013be Rasmus Villemoes 2019-03-07  76   */
6bab69c65013be Rasmus Villemoes 2019-03-07  77  #define static_assert(expr, ...) __static_assert(expr, ##__VA_ARGS__, #expr)

6bab69c65013be Rasmus Villemoes 2019-03-07 @78 #define __static_assert(expr, msg, ...) _Static_assert(expr, msg) 6bab69c65013be Rasmus Villemoes 2019-03-07 79
07a368b3f55a79 Maxim Levitsky 2022-10-25 80

Signature · 2 lines
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
#4 lore raw
In reply to Peter Xu
kernel test robot noticed the following build errors:

Hi Peter,

[This is a private test report for your RFC patch.] kernel test robot noticed the following build errors:

[auto build test ERROR on akpm-mm/mm-everything]

url:    https://github.com/intel-lab-lkp/linux/commits/Peter-Xu/mm-Properly-document-tail-pages-for-compound-pages/20230811-045015
base:   https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git mm-everything
patch link:    https://lore.kernel.org/r/20230810204944.53471-1-peterx%40redhat.com
patch subject: [PATCH RFC] mm: Properly document tail pages for compound pages
config: s390-randconfig-r033-20230811 (https://download.01.org/0day-ci/archive/20230811/[email protected]/config)
compiler: clang version 15.0.7 (https://github.com/llvm/llvm-project.git 8dfdcc7b7bf66834a761bd8de445840ef68e4d1a)
reproduce: (https://download.01.org/0day-ci/archive/20230811/[email protected]/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/

All errors (new ones prefixed by >>):

In file included from arch/s390/kernel/asm-offsets.c:11:
In file included from include/linux/kvm_host.h:16:
In file included from include/linux/mm.h:7:
In file included from include/linux/gfp.h:7:
In file included from include/linux/mmzone.h:22:

>> include/linux/mm_types.h:436:1: error: static assertion failed due to requirement '__builtin_offsetof(struct folio, _flags_2) == __builtin_offsetof(struct page, flags) + 2 * sizeof(struct page)': offsetof(struct folio, _flags_2) == offsetof(struct page, flags) + 2 * sizeof(struct page)
   FOLIO_MATCH(flags, _flags_2);
   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
   include/linux/mm_types.h:434:2: note: expanded from macro 'FOLIO_MATCH'
           static_assert(offsetof(struct folio, fl) ==                     \
           ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   include/linux/build_bug.h:77:34: note: expanded from macro 'static_assert'
   #define static_assert(expr, ...) __static_assert(expr, ##__VA_ARGS__, #expr)
                                    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   include/linux/build_bug.h:78:41: note: expanded from macro '__static_assert'
   #define __static_assert(expr, msg, ...) _Static_assert(expr, msg)
                                           ^              ~~~~
   In file included from arch/s390/kernel/asm-offsets.c:11:
   In file included from include/linux/kvm_host.h:16:
   In file included from include/linux/mm.h:7:
   In file included from include/linux/gfp.h:7:
   In file included from include/linux/mmzone.h:22:
>> include/linux/mm_types.h:437:1: error: static assertion failed due to requirement '__builtin_offsetof(struct folio, _head_2) == __builtin_offsetof(struct page, compound_head) + 2 * sizeof(struct page)': offsetof(struct folio, _head_2) == offsetof(struct page, compound_head) + 2 * sizeof(struct page)
   FOLIO_MATCH(compound_head, _head_2);
   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   include/linux/mm_types.h:434:2: note: expanded from macro 'FOLIO_MATCH'
           static_assert(offsetof(struct folio, fl) ==                     \
           ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   include/linux/build_bug.h:77:34: note: expanded from macro 'static_assert'
   #define static_assert(expr, ...) __static_assert(expr, ##__VA_ARGS__, #expr)
                                    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   include/linux/build_bug.h:78:41: note: expanded from macro '__static_assert'
   #define __static_assert(expr, msg, ...) _Static_assert(expr, msg)
                                           ^              ~~~~
   In file included from arch/s390/kernel/asm-offsets.c:11:
   In file included from include/linux/kvm_host.h:19:
   In file included from include/linux/msi.h:27:
   In file included from include/linux/irq.h:20:
   In file included from include/linux/io.h:13:
   In file included from arch/s390/include/asm/io.h:78:
   include/asm-generic/io.h:547:31: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
           val = __raw_readb(PCI_IOBASE + addr);
                             ~~~~~~~~~~ ^
   include/asm-generic/io.h:560:61: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
           val = __le16_to_cpu((__le16 __force)__raw_readw(PCI_IOBASE + addr));
                                                           ~~~~~~~~~~ ^
   include/uapi/linux/byteorder/big_endian.h:37:59: note: expanded from macro '__le16_to_cpu'
   #define __le16_to_cpu(x) __swab16((__force __u16)(__le16)(x))
                                                             ^
   include/uapi/linux/swab.h:102:54: note: expanded from macro '__swab16'
   #define __swab16(x) (__u16)__builtin_bswap16((__u16)(x))
                                                        ^
   In file included from arch/s390/kernel/asm-offsets.c:11:
   In file included from include/linux/kvm_host.h:19:
   In file included from include/linux/msi.h:27:
   In file included from include/linux/irq.h:20:
   In file included from include/linux/io.h:13:
   In file included from arch/s390/include/asm/io.h:78:
   include/asm-generic/io.h:573:61: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
           val = __le32_to_cpu((__le32 __force)__raw_readl(PCI_IOBASE + addr));
                                                           ~~~~~~~~~~ ^
   include/uapi/linux/byteorder/big_endian.h:35:59: note: expanded from macro '__le32_to_cpu'
   #define __le32_to_cpu(x) __swab32((__force __u32)(__le32)(x))
                                                             ^
   include/uapi/linux/swab.h:115:54: note: expanded from macro '__swab32'
   #define __swab32(x) (__u32)__builtin_bswap32((__u32)(x))
                                                        ^
   In file included from arch/s390/kernel/asm-offsets.c:11:
   In file included from include/linux/kvm_host.h:19:
   In file included from include/linux/msi.h:27:
   In file included from include/linux/irq.h:20:
   In file included from include/linux/io.h:13:
   In file included from arch/s390/include/asm/io.h:78:
   include/asm-generic/io.h:584:33: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
           __raw_writeb(value, PCI_IOBASE + addr);
                               ~~~~~~~~~~ ^
   include/asm-generic/io.h:594:59: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
           __raw_writew((u16 __force)cpu_to_le16(value), PCI_IOBASE + addr);
                                                         ~~~~~~~~~~ ^
   include/asm-generic/io.h:604:59: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
           __raw_writel((u32 __force)cpu_to_le32(value), PCI_IOBASE + addr);
                                                         ~~~~~~~~~~ ^
   include/asm-generic/io.h:692:20: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
           readsb(PCI_IOBASE + addr, buffer, count);
                  ~~~~~~~~~~ ^
   include/asm-generic/io.h:700:20: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
           readsw(PCI_IOBASE + addr, buffer, count);
                  ~~~~~~~~~~ ^
   include/asm-generic/io.h:708:20: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
           readsl(PCI_IOBASE + addr, buffer, count);
                  ~~~~~~~~~~ ^
   include/asm-generic/io.h:717:21: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
           writesb(PCI_IOBASE + addr, buffer, count);
                   ~~~~~~~~~~ ^
   include/asm-generic/io.h:726:21: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
           writesw(PCI_IOBASE + addr, buffer, count);
                   ~~~~~~~~~~ ^
   include/asm-generic/io.h:735:21: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
           writesl(PCI_IOBASE + addr, buffer, count);
                   ~~~~~~~~~~ ^
   In file included from arch/s390/kernel/asm-offsets.c:11:
   In file included from include/linux/kvm_host.h:19:
   In file included from include/linux/msi.h:27:
   In file included from include/linux/irq.h:609:
   In file included from arch/s390/include/asm/hw_irq.h:6:
   In file included from include/linux/pci.h:37:
   In file included from include/linux/device.h:32:
   In file included from include/linux/device/driver.h:21:
   In file included from include/linux/module.h:19:
   In file included from include/linux/elf.h:6:
   In file included from arch/s390/include/asm/elf.h:160:
   include/linux/compat.h:454:22: warning: array index 3 is past the end of the array (which contains 1 element) [-Warray-bounds]
           case 4: v.sig[7] = (set->sig[3] >> 32); v.sig[6] = set->sig[3];
                               ^        ~
   arch/s390/include/asm/signal.h:22:9: note: array 'sig' declared here
           unsigned long sig[_NSIG_WORDS];
           ^

In file included from arch/s390/kernel/asm-offsets.c:11:
In file included from include/linux/kvm_host.h:19:
In file included from include/linux/msi.h:27:
In file included from include/linux/irq.h:609:
In file included from arch/s390/include/asm/hw_irq.h:6:
In file included from include/linux/pci.h:37:

vim +436 include/linux/mm_types.h

7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  412) 
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  413) #define FOLIO_MATCH(pg, fl)						\
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  414) 	static_assert(offsetof(struct page, pg) == offsetof(struct folio, fl))
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  415) FOLIO_MATCH(flags, flags);
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  416) FOLIO_MATCH(lru, lru);
536f4217ced62b Wei Yang                2022-01-29  417  FOLIO_MATCH(mapping, mapping);
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  418) FOLIO_MATCH(compound_head, lru);
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  419) FOLIO_MATCH(index, index);
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  420) FOLIO_MATCH(private, private);
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  421) FOLIO_MATCH(_mapcount, _mapcount);
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  422) FOLIO_MATCH(_refcount, _refcount);
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  423) #ifdef CONFIG_MEMCG
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  424) FOLIO_MATCH(memcg_data, memcg_data);
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  425) #endif
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  426) #undef FOLIO_MATCH
379708ffde1b04 Matthew Wilcox (Oracle  2022-09-02  427) #define FOLIO_MATCH(pg, fl)						\
379708ffde1b04 Matthew Wilcox (Oracle  2022-09-02  428) 	static_assert(offsetof(struct folio, fl) ==			\
379708ffde1b04 Matthew Wilcox (Oracle  2022-09-02  429) 			offsetof(struct page, pg) + sizeof(struct page))
379708ffde1b04 Matthew Wilcox (Oracle  2022-09-02  430) FOLIO_MATCH(flags, _flags_1);
dad6a5eb555648 Hugh Dickins            2022-11-02  431  FOLIO_MATCH(compound_head, _head_1);
379708ffde1b04 Matthew Wilcox (Oracle  2022-09-02  432) #undef FOLIO_MATCH
dad6a5eb555648 Hugh Dickins            2022-11-02  433  #define FOLIO_MATCH(pg, fl)						\
dad6a5eb555648 Hugh Dickins            2022-11-02  434  	static_assert(offsetof(struct folio, fl) ==			\
dad6a5eb555648 Hugh Dickins            2022-11-02  435  			offsetof(struct page, pg) + 2 * sizeof(struct page))
dad6a5eb555648 Hugh Dickins            2022-11-02 @436  FOLIO_MATCH(flags, _flags_2);
dad6a5eb555648 Hugh Dickins            2022-11-02 @437  FOLIO_MATCH(compound_head, _head_2);
dad6a5eb555648 Hugh Dickins            2022-11-02  438  #undef FOLIO_MATCH
7b230db3b8d373 Matthew Wilcox (Oracle  2020-12-06  439) 
Signature · 2 lines
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki